VCloud II, Control and System Engineering in Autonomous drive vehicle with focus on Energy Consumption
|Coordinator||Volvo Personvagnar Aktiebolag - Volvo Car Group R&D|
|Funding from Vinnova||SEK 11 960 000|
|Project duration||March 2015 - December 2018|
|End-of-project report||2014-06251sv.pdf(pdf, 1764 kB) (In Swedish)|
Purpose and goal
When introducing self driving vehicles on the roads, this introduces new type of data and information that can be used to improve the drivetrain efficiency by optimizing the control of the actuators. In VCloudII the knowledge about the future road ahead and destination is used to improve the fuel consumption for a Plug In Hybrid. This introduces the need of finding control methods that can be used on gathered route data and aggregating, isolating and avoiding anomalies in the data, differentiating dynamic disturbances from reoccuring events on the roads.
Expected results and effects
The research emphazises in finding methods and infrastructure to implement AI for self driving vehicles for real time control systems. The aim is to utilize this information when the vehicles are driven and efficiently calculate the optimal control policy for the drivetrain using this data. The result is targeted to be a significant measurable effect of utilizing this new smart control of the drivetrain.
Planned approach and implementation
The work started with a theoretical work package studying the potential of implementing a optimized control method with route data. This was implemented together with the original control software of the vehicle. In the following work packages the method was implemented in a cloud based software architecture together with the existing ECUs in the self driving vehicles. In the last work packages the method was evaluated in real traffic and in a test cell at Chalmers.